ECE398BD: Audio and Visual Analytics (Labs)

Schedule

	Topic	Lab	Quiz	Assigned	Due
Lab 6	Introduction to Audio and Visual Analytics	download	No quiz	Oct 3	Oct 10 @ 23:59
Lab 7	Clustering for Audio and Images	download	Oct 11	Oct 10	Oct 17 @ 23:59
Lab 8	Shazam	download	Oct 18	Oct 17	Oct 24 @ 23:59
Lab 9	Image Retrieval	download	Oct 25	Oct 24	Oct 31 @ 23:59
Lab 10	Working with Videos and Introduction to Deep Learning	download	Nov 1	Oct 31	Nov 7 @ 23:59

Lab 10

You will need to install Tensorflow and run neural networks for this lab. Make sure you do this as early as you can.

Lab 9

You will need to install OpenCV with SIFT for this lab. Make sure you do this as early as you can.

Lab 8

Please use the section of 20-30s for the WakeMeUp.wav
There is a typo in the docstring for function get_2D_peaks. Axis=0 corresponds to frequency, axis=1 corresponds to time.

Lab 7

There is a typo in the function description of ‘calculate_MFCC()‘. The argument M is such that-> M: [int] with (N-M) amount of overlap in consecutive sections

Lab 6

Exercise 3(a): You can pick the first 1500-2000 Hz
Exercise 3(b): For visualizing the Fourier transform of the rectangular and Hamming windows, make sure to pass/set the number of samples ‘n’ of the np.fft.rfft to a number more than the length of the window, e.g. 3*window_length. Play around with the range of frequencies to plot until you get spectrums matching what you saw in ECE310.
Exercise 3(c): Here too, make sure to pass/set the number of samples ‘n’ of the np.fft.rfft to a number more than the length of the window, e.g. 3*window_length.